Misinformation

 

In 1946, a Roper/Fortune survey asked "Here are three different kinds of job:
A job which pays quite a low income, but which you were sure of keeping.
A job which pays a good income but which you have a 50-50 chance of losing.
A job which pays an extremely high income if you make the grade but in which you lose almost everything if you don't make it.
If you had your choice, which would you pick?"

Back in 2013, I had a post on this question, which noted that it was asked again in 1957 and said "unfortunately, the question has never been asked again."  That was wrong--it's been asked twice since then, as well as once in 1948 (maybe I should say "at least," in case I'm still missing some).  The results:

           Safe        50/50      Risky

1946      56%       21%        18%

1948      47%        32%        19%

1957       42%        26%        26%

1962       48%         33%        14%

1981        22%        34%         36%

There is a move towards more risk/more reward, with the exception of 1962.  I'm not sure if that reflects a short-term change or some other difference--the 1962 survey one was done by Gallup while the others were all by the Roper organization.  There was also a difference between 1981 and all previous surveys--the earlier ones had asked men and single women about their job, and asked married women about their husband's job--that is, what you would like him to have.  In 1981, they asked all employed people about their own job (and didn't ask people who weren't employed).  But there was still a change after adjusting for this difference:  "in 1957, men were almost evenly divided among the three options, while almost 50% of women chose the safe job and only about 20% chose the high income/high risk job."  In 1981, only 19% of men chose the safe job and 45% chose the risky one, while women were pretty evenly divided (29%/39%/32%).  

I wondered if opinions on this question were related to political views--that is, people who thought that they could "make the grade" would be more conservative.  The 1981 survey didn't have many political questions, but it did have one about whether you would describe yourself as a supporter or critic of President Reagan.  There was a relationship--people who favored riskier choices were more likely to be supporters, but it became considerably smaller and no longer statistically significant after adding a few controls.  One important control variable was race--black people were more likely to favor the safe job and less likely to support Reagan.  Another was education--more educated people were considerably more likely to favor the risky job, and more likely to support Reagan--here is support for Reagan by education among whites:

not HS grad      62%

HS grad            70% 

some college    75%

bachelor's         80%

graduate           71%

I was a bit surprised by that, even thought I knew that at one time education was associated with support for the Republicans and have written about that in this blog.  I felt like 1981 was part of the "modern" period of political alignments--I don't think that I would have been surprised to see that pattern just a few years before.  I guess the lesson is that thinking in terms of categories or eras has a strong appeal, even when you know that the changes were gradual.  

[Data from the Roper Center for Public Opinion Research]

Too bad to be true

 Michelle Goldberg had a column arguing that there has recently been a turn against feminism.  One piece of evidence is a poll sponsored by the Southern Poverty Law Center.  According to this poll, support for feminism was considerably lower among younger Democrats than among older Democrats.  In Goldberg's summary:

"Predictably, most young Republicans agree with the statement, 'Feminism has done more harm than good.' What was astonishing was how many young Democrats agreed as well. While only 4 percent of Democratic men over 50 thought feminism was harmful, 46 percent of Democratic men under 50 did. Nearly a quarter of Democratic women under 50 agreed, compared with only 10 percent of those 50 and older."  

Here are the figures for all age/party/gender groups.

                 DM             RM                  DW              RW

18-49   41%-46%     31%-62%       71%-23%      34%-52% 

50+      94%-4%       52%-42%       87%-10%      32%-51%

change   -48%              -21%               -15%             +2%

I show the pro-feminist position (disagree) first.  The last row is the change in support for feminism from older to younger, defined as the average of changes in disagree and agree.  For example, among Democratic men, disagree fell by 53 and agree increased by 42, and (53+42)/2 rounds to 48.  So the survey showed a huge drop in support for feminism among Democratic men, substantial drops among Republican men and Democratic women, and essentially no change among Republican women.  

You rarely see numbers like 4% on agree/disagree questions, even among partisans.  For example, in 2012, Gallup asked if "there should or should not be a law that would ban the possessions of handguns, except by police and other authorized persons."  This is an issue that's divided the parties for a long time, and it's hard to imagine a Republican politician supporting it--but 11% of people who identified as Republicans did.  Also, the gap between younger and older Democratic men is implausibly large.  The two groups are just arbitrary divisions in a continuous variable--that is, you're not comparing "young" and "old," but mixed groups that both include a lot of middle-aged people.   

The best match to the SPLC question I could find was in a 2020 Pew survey, which asked "overall, what impact, if any, has feminism had on the lives of" various groups.  The groups included black women, white women, and Hispanic women.  First, I counted "helped" as +1, "hurt" as -1, and "made no difference" as 0.  Then I summed the scores, and counted +3 (helped all groups) as equivalent to disagreeing with the "more harm than good" position, negative scores (hurt more groups than it helped) as agreeing, and scores of 0-2 as mixed.  

The figures for the age/party/gender groups:

                   DM               RM                  DW              RW

18-49      58%-13%      40%-24%       51%-16%      46%-15% 

50+         60%-14%      43%-21%       53%-11%      42%-15%

change       -1%               -3%                  -4%               +2%

They show little generational difference among any group.  Also, although Democrats are more favorable, the partisan differences are smaller than in the SPLC poll.  This seems reasonable to me, since "feminism" can be interpreted in many different ways.  

The SPLC poll shows similarly implausible results on some other questions--for example, only 2% of Democratic men over 50, but 42% of Democratic men aged 18-49 agree that "transgender people are a threat to children."  They didn't give details on their methodology--just said it was from an online panel of 1,500 people--so it's hard to guess what went wrong.  Maybe it was something strange with the weights, so that the group figures are dominated by a few people.  However, that would be more likely to occur in small groups, such as black Republicans, and all of the age/gender/party groups would be expected to be fairly large.  But in any case, the SPLC survey can't be taken seriously.

[Data from the Roper Center for Public Opinion Research]

There's No U

 The New York Times had a story about the health effects of coffee, based on a paper in the Annals of Internal Medicine. The study found that coffee drinkers had substantially lower mortality rates compared to people who didn't drink coffee, even after adjusting for a large number of variables involving health, diet, and exercise.  The point that caught my attention was that the story emphasized that the results involved moderate coffee consumption, and concluded with a warning: 

 "The study showed that the benefits of coffee tapered off for people who drank more than 4.5 cups of coffee each day.  Past studies have shown that consuming 'extreme amounts' — over seven cups per day — can take a toll, she [Christina Wee, professor at Harvard medical school and deputy editor of the Annals] said.

'Moderation is good,' Dr. Goldberg [professor at NYU medical school] said. 'But too much of a good thing isn’t necessarily more of a good thing.'"  

I like coffee, and probably fall into the "extreme amounts" group, so I wanted to see how strong the evidence was.  

The estimated relationship between coffee consumption and mortality rates is shown in this figure from the paper:

The first column is people who drink unsweetened coffee, the second people who drink coffee with sugar, and the third is people who drink it with artificial sweetener--all are compared to people who didn't drink coffee.  For unsweetened coffee, it looks like the reduction in mortality is about the same for all of the range of 3-8 cups, and for coffee with artificial sweeteners, it gets larger the more you drink.   It's only for coffee with sugar that there's evidence that high amounts are worse than moderate amounts.  The confidence intervals are pretty wide in all cases, so they don't rule out the possibility that the reduction in mortality gets smaller with larger amounts of unsweetened coffee, but they don't support it either.   It seemed strange that the Times story emphasized a point that wasn't really supported by the analysis, but on reading the article more closely, I can see why it did.  It was just following the paper, which mentioned a "U-shaped association with mortality" several times.  Moreover, the paper was accompanied by an editorial (written by Dr Wee) which said "they found a U-shaped association."  

The main focus of the paper wasn't the shape of the relationship, but whether there was a difference between sweetened and unsweetened coffee--specifically, whether the (apparent) benefits were smaller for sweetened coffee.  In fact, the sample wasn't very useful for estimating the shape, because there weren't many people who drank large amounts.  The paper didn't describe the distribution in detail, but the mean number of cups per day for those who drank coffee was 2.4, with a standard deviation of 1.3, so 7 cups would be about 3.5 standard deviations above the mean.*  

So why did the paper and the editorial emphasize a secondary point and mischaracterize it?  I think it's a combination of two things:  

1.  Traditionally people modeled non-linear relationships using polynomial regressions, starting (and often ending) by adding x squared.  A regression with x and x-squared will always be symmetrical around the minimum/maximum.  So people got into the habit of  thinking in terms of U or inverted U shapes, and that has persisted even after more flexible methods of modelling non-linear relationships have come into use.  

2.  Medical people are inclined to start from the assumption that extreme consumption of anything is bad.

2a.  When communicating with the public, they are not purely concerned with conveying information, but also with getting people to do the right thing, or not encouraging them to do the wrong thing.  Since some people tend to excess, emphasizing moderation is part of that.

*The data were from Britain.

Who cares about gun control?

 My last post discussed a question about the importance of the abortion issue in voting:  "if you would only vote for a candidate who shares your views on abortion, consider a candidate's position on abortion as just one of many important factors, or not see abortion as a major issue."  A parallel question has been asked about gun control.  There was also one on immigration policy, although it was asked only a couple of times over a short period.  The means, with higher numbers meaning more important:

The self-rated importance of both abortion and gun control has increased in the 21st century.  I interpret that as one aspect of a general increase in ideological polarization.  A more surprising point is that gun control consistently ranks above abortion in importance (and immigration does as well on the two occasions when it was asked)--my impression, which I think is widely shared, was that abortion was the most important "social issue," and gun control was one of a number of issues in the second rank.  The difference in means is primarily a result of differences in the number who say it's not a major issue.  

Maybe that's because some people didn't think that elected officials had that much influence over abortion--they thought that the courts would decide.  

In my last post, I found that people who were opposed to legal abortion rated it as more important than those who were in favor.  There was also a tendency for people with middle opinions (e. g., should be allowed only in the first trimester) to rate it as less important.  A CNN/ORC poll from January 2013 included the question on the importance and a number of questions about gun control measures.  For a basic one, "do you favor or oppose stricter gun control laws" people who were opposed were more likely to say that they would only vote for a candidate who shares their views (19% vs. 15%) but also more likely to say it wasn't a major issue (20% to 14%).   The means were almost identical.  A majority favored stricter laws, but it wasn't overwhelming (about 55%).  There was also a three-way question:  "do existing laws make it too easy to buy guns, too difficult, or are they about right?"  A majority (52%) of the people who said "too difficult" said that they would only vote for someone who shares there views; however, only a tiny minority, about 3%, thought it was too difficult.  There was little or no difference between ratings of importance in the "too easy" or "about right" groups.  There was a question about a ban on "semi-automatic assault guns"--people who were opposed were more likely to rate the issue as essential, but also more likely to rate it as not important.  There was also a question on whether you had a gun in your household--that also showed little or no difference in ratings of the importance of the issue.  

So far, this doesn't give much support to the conventional view that opponents of gun control are more passionate about the issue.  But there was also a question, asked of gun owners, about whether you felt that the federal government was taking away your right to own a gun.  Among gun owners who said yes, 24% said they would only vote for someone who shared their views and 14% said it wasn't a major issue; among gun owners who said no, the figures were almost exactly reversed, with 12% saying essential and 24% not a major issue.  Gun owners who didn't think the federal government was trying to take away their right to own a gun rated the issue as less important than non-owners.  

I think this last point says something about party competition today--rather than supporting measures that voters in the middle want, parties try to convince their supporters that they are the think end of the wedge:  e. g., an apparently reasonable restriction on gun ownership is just one step in a plan to take away your guns.  That might help to explain a paradox discussed in the New York Times today--that gun control measures which seem popular in the polls often get much less support when they are offered to voters in state referendums.

[Data from the Roper Center for Public Opinion Research]